Skip to content
Sphere · Whitepaper

Trust-Based Social Network

Modern social networks today have largely transformed into “social media”, with emphasis on the usage of media. While modern computing offers enormous new capabilities for exchanging information through the internet, various social and economic factors have incentivized rapid development towards media and entertainment while neglecting advancements in other areas.

The trust-based social network is a decentralized social network that aims to better represent the network structure of human interaction and provide a scalable and flexible, yet stable platform for human organisation.

Modelling Human Interaction

Section titled “Modelling Human Interaction”

Human interaction is a very complex phenomenon lying at the cross-section of many disciplines, including game theory, network theory, psychology, neurobiology, politics, economics and even spirituality.

All social networks aim to represent some kind of human interaction. Facebook’s “friend system” is a one-dimensional projection: you are either friends or you are not. Instagram and YouTube shift to “followers” and “subscribers.” These are relatively one-dimensional projections of the complex interactions between humans.

In real life, a relationship can have multiple different statuses and descriptors. A large portion of any relationship is related to some kind of trust, often regarding different things on different levels. Person A might trust Person B to see some of their pictures, but might not want to allow them to call at any time — providing an Instagram handle instead of a phone number. You can think of the account handle or the phone number as unique keys that represent different privileges.

Basic P2P Trust Model

Section titled “Basic P2P Trust Model”

The goal of the underlying software network is to establish and maintain an abstractly represented social network graph of connections between individuals.

Entities can represent users but also groups, roles, or other virtual actors like automated algorithms. They have certain attributes (e.g. “name”, “age”) and directional connections of different types (e.g. “knows”, “trusts”).

Connections carry information and serve different functions depending on their type, but generally denote a relationship between two entities. A basic connection is initially directional as it is initiated by a person. It stays directional until the receiver reciprocates with a connection of the same type.

Groups, called "Spheres", are entities representing groups of individuals which can then act as individual entities in the network.

Permission System

Section titled “Permission System”

The basic principles of privacy should allow very fine-grained control over one’s own data, but also control over the way information is shared and received.

Basic Ideas:

  • A person may only establish a connection from themselves to others and may not force establishment of a connection towards them
  • A person may object to a connection established towards them, causing the connection to be removed/blocked
  • There should be a highly customizable, hierarchical permission system controlling which information is known to which people

Identity

Section titled “Identity”

See Identity for the detailed treatment of identity challenges including 1-to-1 identity, identity recovery, and the n-to-m identity problem.

Web3 and Blockchain Base Layers

Section titled “Web3 and Blockchain Base Layers”

Bitcoin and other web3 technologies have created a large infrastructure of cryptographic, decentralized networks. Many are programmable with smart contracts, creating custom currencies, tokens, voting systems, financial platforms and DAOs.

Trust-based social networks take a different structural approach, but can benefit immensely from integrations with existing blockchain ecosystems. Blockchains and smart contracts can be used to build up parts of the system, providing a somewhat standardized computational layer with high-level network features.

Key Differences to Current Models

Section titled “Key Differences to Current Models”
  • Focus on networking between people, not media consumption
  • Discourage unnecessary transactions by fees (not a currency network, but has internal currency for paying network providers)
  • User-controlled data sharing and algorithmic transparency (Web3 Social Media)

Network Theory

Section titled “Network Theory”

Decentralization can be measured using “Centrality Indicators” which describe particular nodes in terms of how well-connected and integrated a node is. Four important centrality measures: degree, betweenness, closeness and eigenvector. The “mean decentrality” of a network provides insights into its overall distribution of power.

Pareto’s Principle (the 80-20 rule) is the natural tendency in networks — 80% of consequences caused by 20% of factors. While this can be leveraged, it might also need to be limited through structural setups to prevent excessive centralization.

Section titled “Related Concepts”